Claude Code has two billing systems. A subscription includes a usage pool that resets over time, while API access charges for every token processed. The coding agent is the same; usage tracking, costs, and limits are different.
Pro, Max, and Team bundle Claude Code usage into the subscription price. Five-hour and weekly limits apply, and activity in other Claude products draws from the same allowance. API billing has no included allowance. It charges by token until the account reaches a credit, spending, or API rate limit.
Choose Pro for individual repository work, Max for daily high-volume sessions, Team or Enterprise for organization-managed access, and API billing for automation or project-level cost tracking.
Table Of Contents
- Claude Code pricing at a glance
- Can Claude Free use Claude Code?
- Claude Pro pricing and limits
- Claude Max 5x and Max 20x
- Team Standard and Premium seats
- Enterprise pricing and usage
- Claude Code in the terminal, IDE, and web
- How the five-hour limit works
- How weekly limits work
- What consumes Claude Code usage
- Model choice, plan access, and usage
- Extra Usage and usage credits
- Claude Code API pricing
- Claude subscription login versus an API key
- What happens when a limit is reached
- How to check current usage
- Which plan fits each workload?
- Reducing Claude Code usage without reducing useful work
Claude Code pricing at a glance
All prices are in US dollars before tax. Annual prices require payment for the full year.
| Plan | Price | Usage |
|---|---|---|
| Free | $0 | Claude chat only |
| Pro | $20/month or $200/year | Included Claude Code limits |
| Max 5x | $100/month | 5x Pro session usage |
| Max 20x | $200/month | 20x Pro session usage |
| Team Standard | $25/seat/month or $20/seat/month with annual billing | More usage than Pro |
| Team Premium | $125/seat/month or $100/seat/month with annual billing | 5x Standard usage |
| Enterprise | $20/seat/month plus API usage | Organization-managed billing |
| API or Console | Pay by token | No included subscription usage |
Can Claude Free use Claude Code?
Claude Free does not include Claude Code in the terminal, IDE extensions, or Claude Code on the web. The Free plan generates code, analyzes files, creates files, and executes code inside Claude chat. That work stays inside the Claude app and does not give the coding agent access to a local repository.
Claude Code works inside a repository. It inspects project files, edits code, runs commands and tests, uses Git, and maintains a coding session across many actions. The terminal and supported IDEs require a paid Claude subscription or a Claude Console account. Claude Code on the web is available to Pro, Max, Team, and Enterprise accounts with Claude Code access.
A free Claude account is enough to test Claude’s answers about code. To let the coding agent work directly in a local project, use Pro or another paid plan. A Console account provides the same local access and bills every request at API rates.
Claude Pro pricing and limits
Claude Pro costs $20 per month. The annual option costs $200 upfront, which equals $16.67 per month when divided across the year. One subscription includes Claude chat and Claude Code under the same account.
Pro has no fixed prompt count. A short request that reads one file costs much less of the allowance than a debugging session that scans a repository, runs tests several times, and carries a long conversation. The selected model and effort level also affect consumption.
Pro usage is shared across Claude. Messages in the web, desktop, and mobile apps draw from the same pool as Claude Code. A large chat before a coding session leaves less capacity for the terminal in the same window.
Claude Max 5x and Max 20x
Max is an individual subscription for people who use Claude throughout the day. It has two monthly tiers:
- Max 5x costs $100 per month and provides five times the Pro usage allowance in each five-hour session.
- Max 20x costs $200 per month and provides twenty times the Pro usage allowance in each five-hour session.
The multiplier describes session usage compared with Pro. It does not remove weekly limits, model caps, or other capacity controls. Max accounts have an all-model weekly limit and a separate Sonnet weekly limit. The account’s Usage page shows the assigned reset time.
Max 5x supports regular feature work, repeated debugging sessions, and developers who exhaust Pro. Max 20x supports sustained daily work, several active repositories, and frequent parallel sessions. Neither tier is unlimited. Fable sessions, several agents, and multiple Claude Code processes consume Max capacity faster.
Team Standard and Premium seats
The Team plan supports organizations with 2 to 150 members. Standard and Premium seats share one organization, and owners assign the required allowance to each member.
A Standard seat costs $25 per month or $20 per month with annual billing. It includes more usage than Pro and supports occasional or moderate coding work.
A Premium seat costs $125 per month or $100 per month with annual billing. It has five times the usage of a Standard seat. Teams assign Premium seats to developers who spend much of the day in Claude Code and Standard seats to lighter users.
Team usage belongs to the organization account. Owners manage seats, billing, connectors, and Claude Code settings. The plan also adds central administration, single sign-on, usage analytics, and a default policy that excludes customer content from model training.
Team members should confirm that Claude Code is using the organization login. A personal login draws from a personal subscription. An API key in the environment sends usage to the Console bill.
Enterprise pricing and usage
Enterprise currently costs $20 per seat per month plus usage at API rates and is billed annually. The seat provides access to the product and administrative controls. Model charges change with the amount of work performed.
Pro, Max, and Team plans include an allowance. Usage-based Enterprise accounts charge for tokens from the first request and have no included five-hour pool.
Anthropic also supports seat-based Enterprise arrangements. Standard and Premium seats on those plans have included usage. An owner enables usage credits after a member reaches the seat limit. The company’s contract and organization settings define the billing system.
Enterprise adds role-based permissions, user and organization spend limits, audit logs, the Compliance API, custom retention settings, network controls, and usage reporting.
Claude Code in the terminal, IDE, and web
The terminal, IDE extensions, and web service use the same core agent. Each runs the work in a different environment.
| Interface | Runs on | Access and billing |
|---|---|---|
| Terminal | Local machine | Paid plan or API account |
| VS Code and Cursor | Local IDE | Paid plan or API account |
| JetBrains | Local IDE | Paid plan or API account |
| Claude Code web | Anthropic cloud | Pro, Max, Team, or Enterprise |
The terminal is the most direct local interface. It reads the repository, edits files, runs shell commands, and uses local development tools. The IDE integrations place the same coding workflow beside the editor, with inline diffs, file references, plan review, and conversation history.
Claude Code on the web runs in an isolated cloud environment. It connects to a repository, continues after the browser closes, and prepares a branch or pull request. Web and local tasks draw from the same account capacity.
Related: Claude Code Commands Cheat Sheet
How the five-hour limit works
Claude subscription plans use a rolling five-hour session window. The window measures account consumption over time and does not guarantee five continuous coding hours. Demanding tasks exhaust the allowance faster than short, focused requests.
Usage becomes available again when the window resets. Claude shows the reset time when the account approaches or reaches its limit. Reset times do not follow midnight. Unused capacity does not accumulate for a later window.
Activity across Claude products counts toward the same subscription capacity. Claude chat on the web, desktop, or mobile reduces the capacity available to Claude Code in the same window. Terminal, IDE, and web coding sessions also share the account pool.
Anthropic does not publish a dependable message count for each plan. A fixed number would be misleading because token consumption changes with the task, model, and context. Your account’s Usage page shows your assigned allowance.
How weekly limits work
Paid plans add weekly controls on top of the five-hour window. A five-hour reset does not reset the weekly allowance. Fresh session capacity does not restore weekly capacity.
Max plans have two visible weekly limits. One applies across all models, and the other applies to Sonnet. The reset day and time are assigned to the account and stay fixed from week to week. The Usage page displays every active weekly, model, and feature limit for the account.
Weekly limits prevent a high session allowance from becoming unlimited continuous use. Several heavy days exhaust the weekly pool even when each session stays under its five-hour limit.
What consumes Claude Code usage
Tokens sent to and returned by the model determine Claude Code usage. Each coding turn includes the conversation history, project instructions, file contents, tool definitions, command output, error logs, and summaries from other agents.
The following patterns raise consumption:
- Long conversations resend or summarize more history on later turns.
- Large repositories require more file reading and search results.
- Fable and Opus consume more of a subscription allowance than lighter models.
- Higher effort settings spend more tokens on reasoning.
- Large test logs, generated files, and pasted stack traces add context.
- MCP servers add tool definitions before the model calls any tool.
- Subagents and agent teams create separate model conversations.
- Parallel Claude Code sessions draw from the same account at the same time.
- Repeated corrections keep old failed approaches in the conversation context.
Context window limits are separate from usage limits. The context window controls how much information a model considers in one conversation. The usage limit controls how much model work the account performs over time. A conversation still has context space after the account exhausts its five-hour capacity.
Use /clear when a new task does not need the old conversation. Use /compact when the current task needs a shorter summary of earlier work. These commands reduce future token use. Capacity already consumed stays charged to the active limit window.
Model choice, plan access, and usage
Run /model inside Claude Code to see the models assigned to the signed-in account. The list reflects the account plan, organization policy, and current product availability.
| Model | Best fit and plan access |
|---|---|
| Fable 5 | Hardest long-running agents and complex multi-stage work. Pro uses credits; Max 5x and Max 20x include 50% of the weekly limit |
| Opus 4.8 | Complex coding and agent work at half Fable’s API token price |
| Sonnet 5 | Daily feature work, debugging, tests, and reviews |
| Haiku 4.5 | Fast searches, small edits, and mechanical tasks |
Fable 5 is the first choice for the hardest long-running and multi-stage tasks. Sonnet 5 handles daily coding at lower cost. Opus 4.8 remains a complex-coding model at half Fable’s input and output price. Haiku 4.5 suits short tasks that do not require deep reasoning.
Changing models does not clear the conversation or restore used capacity.
Model versions change more frequently than subscription names. Check the release history when an older model name appears in account settings or documentation.
Related: Claude Code Timeline: Release Date and Major Updates
Extra Usage and usage credits
Anthropic now calls the paid overage system “usage credits.” Some account screens and older material use “Extra Usage.” Both names refer to paid consumption after the included subscription allowance runs out.
On Pro, Max 5x, and Max 20x, the account uses included capacity first. Claude shows a warning when that capacity is exhausted. Enabled and funded usage credits continue the session at standard API rates.
Individual subscribers manage credits in Settings > Usage. The controls include prepaid funds, monthly spending limits, usage alerts, and auto-reload. Auto-reload adds funds when the balance falls below the chosen threshold. Usage-credit charges are separate from the subscription price.
Team owners buy credits for the organization and set organization-wide or user-level spending limits. Only an owner enables the feature. Seat-based Enterprise controls apply at the organization, seat type, group, or user level. The organization’s Usage page shows its available controls.
Usage credits do not change the normal reset schedule. The included five-hour pool resets at its assigned time. Paid credit consumption applies after the included pool is exhausted.
Claude Code API pricing
API billing charges for input tokens, output tokens, and prompt-cache operations. One MTok means one million tokens. As of July 23, 2026, the first-party Claude API uses the following rates.
| Model | Input per MTok | Output per MTok |
|---|---|---|
| Claude Fable 5 | $10.00 | $50.00 |
| Claude Opus 4.8 | $5.00 | $25.00 |
| Claude Sonnet 5 | $2.00 | $10.00 |
| Claude Haiku 4.5 | $1.00 | $5.00 |
Prompt caching prices
| Model | Cache write per MTok | Cache read per MTok |
|---|---|---|
| Claude Fable 5 | $12.50 | $1.00 |
| Claude Opus 4.8 | $6.25 | $0.50 |
| Claude Sonnet 5 | $2.50 | $0.20 |
| Claude Haiku 4.5 | $1.25 | $0.10 |
Sonnet 5 has introductory pricing through August 31, 2026. Its standard price from September 1, 2026 is $3 per million input tokens, $15 per million output tokens, $3.75 per million cache-write tokens, and $0.30 per million cache-read tokens.
Compare model rates: AI LLM API Pricing
Some tools have separate charges. Web search, code execution beyond included allowances, managed agent runtime, and cloud-provider features use their published rates. Regional inference and fast model modes apply their listed multipliers.
Prompt caching lowers the cost of repeatedly reading stable context. Claude Code reuses cached project instructions and other repeated prompt content. Cache writes have their own price, and cache reads cost less than sending the same tokens as new input.
API billing also has rate limits. These control requests or tokens over short periods and are separate from the five-hour and weekly subscription limits. A traffic spike triggers a rate-limit error even when the API account has credit.
Claude subscription login versus an API key
Claude Code decides how to bill usage from the authentication method it is using.
With subscription authentication, /login connects Claude Code to a Pro, Max, Team, or Enterprise account. Interactive usage draws from that plan or seat. The subscription price does not create a Claude API credit balance.
With API authentication, Claude Code sends requests through a Console account, Amazon Bedrock, Google Cloud, Microsoft Foundry, or another supported provider. The provider bills by token under its own prices, rate limits, and spending controls.
When ANTHROPIC_API_KEY is set, Claude Code uses that key in place of the subscription login. The work creates API charges even if the user also pays for Pro or Max.
Check the active account after changing login methods or development environments. Claude Code inherits keys from shell profiles, IDE configurations, containers, and CI secrets even when the current command does not display them.
Subscription and API balances stay separate. Pro includes Claude Code usage within its plan limits. Console credit pays API charges and does not increase the Pro allowance. Usage credits are the paid extension to a subscription limit.
What happens when a limit is reached
| Limit | Result and next step |
|---|---|
| Five-hour subscription | Included use stops until the displayed reset time. Wait, enable credits, switch to billed API usage, or upgrade |
| Weekly subscription | Included use stops until the assigned weekly reset. Wait, use credits, or change plans |
| Model-specific | Select another listed model or wait for its reset |
| Usage-credit balance | Add prepaid funds or enable auto-reload |
| Monthly spend | The owner raises the limit or paid use waits for the next billing period |
| API rate | Reduce concurrency and retry with backoff, or request a higher tier |
| API credit or billing | Add credit, correct billing, or raise the approved spending limit |
| Context window | Run /compact, start a new session, or load less material |
Team and Enterprise members request additional capacity from an administrator. The owner enables usage credits, raises a user limit, or assigns a seat with more usage.
How to check current usage
Individual subscribers open Settings > Usage in Claude. This page shows the current five-hour and weekly bars, reset times, usage credits, and month-to-date credit spending. It includes activity across Claude products and provides a more complete view than a local terminal record.
Inside Claude Code, /usage shows plan usage bars, recent activity, and a breakdown of local consumption. It identifies models, skills, subagents, plugins, or MCP servers that account for a large share of recent use. The local session figures do not include work from another machine or the Claude web app.
The /status command confirms the active account, model, and configuration. Use it after switching between subscription and Console authentication. Check the Claude Usage page or Console for billing totals.
The /usage-credits command opens credit controls for a Claude subscription login. It is not available when Claude Code is authenticated only with an API key.
If you use API billing, check the Claude Console Usage page for the final billing record. Claude Code estimates session cost from local token counts. The Console also includes server-side calculations, adjustments, and usage from every client that shares the account.
Related tool: Claude Code Usage Monitor
Team and Enterprise owners have organization analytics for adoption, activity, and spend. Member-level views depend on the organization’s settings and the viewer’s role.
Which plan fits each workload?
Start with the least expensive option that supports the normal week. Check the Usage page after several real projects. Repeated interruptions show that the plan is too small. Exclude one unusually large repository or debugging session from the comparison.
| Workload | Recommended option |
|---|---|
| Code explanations and generated snippets in Claude chat | Free |
| A few focused repository tasks each week | Pro |
| Regular feature work and debugging | Max 5x |
| Heavy daily coding across several repositories | Max 20x |
| A team with mixed usage levels | Team with Standard and Premium seats |
| A development team with frequent Claude Code use | Team Premium |
| Security, identity, audit, and policy requirements | Enterprise |
| CI jobs, scripts, Agent SDK work, or unattended automation | API or separate Agent SDK credits |
| A short period of unusually heavy coding | Pro or Max with usage credits |
| Project budgets and machine-driven workloads | API with workspace limits |
Reducing Claude Code usage without reducing useful work
- Start a new session when moving to an unrelated task.
- Ask Claude to inspect named files or functions before searching the entire repository.
- Keep generated files, build output, and large logs out of the working context unless they contain the error.
- Use Sonnet for routine implementation and reserve costlier models for difficult decisions.
- Run
/compactbefore a long conversation becomes difficult to manage. - Disable unused MCP servers. Their tool definitions enter every request.
- Limit parallel agents to independent tasks.
- Give a clear target, constraints, and acceptance test in the first prompt.
FAQs
Do weekly limits reset when a new billing month starts?
No. Weekly limits follow the reset day and time assigned to the account. The subscription renewal date and weekly reset schedule are separate.
Can usage credits be turned off?
Yes. Individual subscribers disable usage credits in Settings > Usage. Claude then stops paid overage when the included plan allowance runs out. Team and Enterprise credit settings are controlled by organization owners.
Does switching models reset a usage limit?
No. A model change affects future consumption within the same session and weekly pools. It does not restore capacity already used.
Does upgrading from Claude Pro to Max take effect immediately?
Yes. The Max plan starts immediately after the upgrade. Anthropic applies the unused value of the current subscription to the new plan.
Does canceling Pro or Max remove Claude Code immediately?
No. The paid plan remains active until the end of the current billing period. The account then moves to Free, which does not include Claude Code access.
Does claude -p use the normal Claude Code subscription limit?
No. Since June 15, 2026, claude -p and the Claude Agent SDK use a separate monthly credit on eligible paid plans. The credit is $20 for Pro and Team Standard, $100 for Max 5x and Team Premium, $200 for Max 20x and seat-based Enterprise Premium, and $20 for usage-based Enterprise. Seat-based Enterprise Standard does not receive this credit.
Does Claude Code on the web charge for the cloud virtual machine?
No. Anthropic does not add a separate virtual-machine compute charge for Claude Code on the web. The task consumes the signed-in account’s Claude Code allowance.
Does Anthropic train models on Claude Code sessions?
For Free, Pro, and Max accounts, model-training use follows the user’s privacy setting. Team, Enterprise, and API data is not used for model training by default.
Related resources
- The Ultimate Claude Code Resource List: Agents, Skills, Plugins & More
- 7 Best CLI AI Coding Agents (Open Source)
- Claude Code Local: Run Claude Code With Local AI on Apple Silicon
